写一个Matlab从配送中心出发，连接其它点，并使路线长度之和最短

您的问题涉及到的是一个典型的旅行商问题（Traveling Salesman Problem，TSP），是一个NP难问题，通常使用启发式算法求解。以下是一个使用遗传算法（Genetic Algorithm，GA）求解TSP的Matlab代码示例：

% 配置参数
popSize = 50; % 种群大小
numCities = 20; % 城市数
maxGen = 500; % 进化代数
pc = 0.8; % 交叉概率
pm = 0.01; % 变异概率

% 生成随机城市坐标
cities = rand(numCities, 2) * 100;

% 初始化种群
pop = zeros(numCities, popSize);
for i = 1:popSize
    pop(:, i) = randperm(numCities);
end

% 进化过程
for gen = 1:maxGen
    % 计算适应度值
    dist = zeros(popSize, 1);
    for i = 1:popSize
        dist(i) = sum(sqrt(sum((cities(pop(:, i), :) - circshift(cities(pop(:, i), :), [1, 0])).^2, 2)));
    end
    fitness = 1./dist;
    fitness = fitness./sum(fitness);

    % 种群选择
    [val, idx] = sort(fitness, 'descend');
    elite = pop(:, idx(1));
    pop = pop(:, idx(1:popSize-1));
    prob = cumsum(val);
    prob = prob/prob(end);
    for i = 1:popSize-1
        a = find(rand() <= prob, 1);
        b = find(rand() <= prob, 1);
        if rand() <= pc
            % 交叉
            c = randi(numCities-1);
            d = randi([c+1, numCities]);
            child = [pop(1:c, a); pop(c+1:d, b); pop(d+1:end, a)];
            child = unique(child);
            if length(child) < numCities
                % 补全缺失的城市
                missing = setdiff(1:numCities, child);
                child = [child; missing];
            end
            pop(:, i) = child;
        else
            % 复制
            pop(:, i) = pop(:, a);
        end
    end

    % 变异
    for i = 2:popSize-1
        if rand() <= pm
            a = randi(numCities);
            b = randi(numCities);
            pop([a, b], i) = pop([b, a], i);
        end
    end
    
    % 加入精英
    pop(:, end) = elite;
end

% 输出最优解
dist = zeros(popSize, 1);
for i = 1:popSize
    dist(i) = sum(sqrt(sum((cities(pop(:, i), :) - circshift(cities(pop(:, i), :), [1, 0])).^2, 2)));
end
[val, idx] = min(dist);
fprintf('最短路程长度：%f\n', val);
fprintf('最短路程路径：');
fprintf('%d -> ', pop(:, idx));
fprintf('%d\n', pop(1, idx));

该代码使用遗传算法求解TSP问题，将城市坐标随机生成，通过进化过程逐步优化种群中的路线。最终输出最短路程长度和路径。您可以根据实际情况进行参数调整和优化。